-1

Showing $\mathbf Q(\sqrt2,\sqrt3)=\mathbf Q(\sqrt2+\sqrt3)$

BUT I want to show this using The Theorem of the Primitive Element,

enter image description here

So I have to verify that $c$ cannot be $1$ and I need the $\mathbf Q$-monomorphisms, what are these ?

user257
  • 989
  • @ccorn the same question yes, but I want a different answer – user257 Aug 11 '16 at 16:56
  • 2
    The primitive element theorem does not help at all in demonstrating that one particular element is a primitive element; the theorem just says that at least one primitive element exists. – Zev Chonoles Aug 11 '16 at 17:01

1 Answers1

-1

Well, there's a mistake in my argument which is detailed in the comment section.

I think it is beneficial to leave this answer as is in case somebody comes upon this and can then learn from it too.

My original post:

Here's one way to do it:

Theorem

Let $F$ be an infinite field. If $F(\alpha_1,\alpha_2)$ is a simple and algebraic extension of $F$ then $\exists t \in F$ s.t $F(\alpha_1,\alpha_2) = F(\alpha_1 + t\alpha_2)$.

This can be generalized for any n BTW

Proof:

I'll assume you know (since we're using the primitive element theorem) that under these conditions there is only a finite number of sub-fields in $F$.

$F$ is infinite then $\exists t_1 \neq t_2 \in F$ s.t $F(\alpha_1 + t_1\alpha_2) = F(\alpha_1 + t_2\alpha_2)$.

Then $\alpha_2 = \frac{(\alpha_1 + t_1\alpha_2)-(\alpha_1 + t_2\alpha_2)}{t_1-t_2} \in F(\alpha_1 + t_1\alpha_2)$ then clearly $\alpha_1 \in F(\alpha_1 + t_1\alpha_2)$ as well.

This ends that theorem.

Now $\mathbb{Q}(\sqrt{2},\sqrt{3})$ applies to the above by the primitive root theorem.

Then $\mathbb{Q}(\sqrt{2} + \sqrt{3}) \subset \mathbb{Q}(\sqrt{2},\sqrt{3}) = \mathbb{Q}(\sqrt{2}+ t\sqrt{3})$

Now notice that $t \neq 0$ necessarily and that $t\sqrt{2} + t\sqrt{3} = a_0 + a_1\sqrt{2} + a_1t\sqrt{3} $

implying $ 0 = a_0 + (a_1-t)\sqrt{2} + (a_1t-t)\sqrt{3} $ and as $\{a_0,\sqrt{2},\sqrt{3}\}$ is a linearly independent set we have $a_1 = t = a_1t$ then $t=1$

Mariah
  • 3,760
  • Your argument is incorrect - you cannot conclude from $\mathbb{Q}(\sqrt{2}+\sqrt{3})=\mathbb{Q}(\sqrt{2}+t\sqrt{3})$ that $t=1$. The step where you say that necessarily $$t\sqrt{2}+t\sqrt{3}=a_0+a_1\sqrt{2}+a_1\sqrt{3}$$ is your mistake. – Zev Chonoles Aug 11 '16 at 17:05
  • @ZevChonoles thanks for the comment, I was wondering that myself: doesn't $Q(\sqrt2 + t\sqrt3) = {a + b(\sqrt2 + t\sqrt3) | a,b \in Q}$? that would imply my argument, no? – Mariah Aug 11 '16 at 17:08
  • btw, downvoting an answer with a mistake isn't the greatest incentive to attempt stuff.. comments will clarify what is necessary in my opinion.. – Mariah Aug 11 '16 at 17:10
  • That's not correct; for convenience, let $\alpha=\sqrt{2}+t\sqrt{3}$, then since $[\mathbb{Q}(\alpha):\mathbb{Q}]=4$, we have $$\mathbb{Q}(\alpha) ={a+b\alpha+ c\alpha^2+d\alpha^3:a,b,c,d\in\mathbb{Q}}$$ – Zev Chonoles Aug 11 '16 at 17:11
  • I made a minor formatting edit to your answer so that the Stack Exchange system would permit me to remove my downvote. – Zev Chonoles Aug 11 '16 at 17:16
  • Ah indeed, glad I made that mistake then, since that is a very basic error to make in this subject. Thanks again, and for removing the downvote. – Mariah Aug 11 '16 at 17:21